INFECTION AND IMMUNITY, Sept. 1990, p. 2951-29560019-9567/90/092951-06$02.00/0Copyright C 1990, American Society for Microbiology

Binding of Mycobacterium avium-Mycobacterium intracellulare toHuman Leukocytes

ANTONINO CATANZARO1* AND SAMUEL D. WRIGHT2

Department of Medicine, Division of Pulmonary and Critical Care Medicine, University of California, San Diego,San Diego, California 92103,' and Laboratory of Cellular Physiology and Immunology,

The Rockefeller University, New York, New York 100212

Received 12 October 1989/Accepted 11 June 1990

We examined nonopsonic binding of Mycobacterium avium-Mycobacterium intracellulare (MAI) by humanleukocytes. Macrophages (M4) avidly bound fluorescently labeled MAI in the absence of serum proteins.Binding appeared to be mediated by a lineage-specific, proteinaceous receptor on M4, since (i) binding oflabeled bacteria could be competitively inhibited by unlabeled MAI, (ii) treatment of M4 with trypsin ablatedthe ability of M+ to bind MAI, and (iii) the capacity to bind MAI was observed on monocytes, M4, andstimulated polymorphonuclear cells but not on lymphocytes or unstimulated polymorphonuclear cells. Thereceptor for MAI appeared mobile in the plane of the membrane, since spreading of M4 on a carpet ofimmobilized, unlabeled MAI down modulated binding of labeled MAI added in suspension. The recep-

tor required neither calcium nor magnesium for activity and appeared different from other known receptorsfor intracellular pathogens.

The soil and water in large regions of the United States are

contaminated with Mycobacterium avium-Mycobacteriumintracellulare (MAI) (1). Epidemiologic studies in theseregions by using delayed-type hypersensitivity to antigensderived from MAI suggest that infection with MAI is very

common (15). Despite repeated contact with this organism,clinical infection among the general population is very

infrequent (15). However, when host defenses are altered asin patients with chronic lung diseases (10) or under condi-tions that depress cell-mediated immune responses, thebalance shifts and chronic or disseminated (16) diseasedevelops. This is particularly the case for acquired immuno-deficiency syndrome (AIDS), where overwhelming infectionwith MAI is found in more than half of the victims (7, 25).MAI is an intracellular pathogen that resides principally in

the macrophage (MO), as is apparent from the examinationof infected tissues of both AIDS and non-AIDS patients (24).In vitro studies demonstrate that M4X provide an excellentenvironment for replication of MAI (3, 5, 9). An intracellularlocation may help MAI evade host defense mechanisms.Here we examine the earliest interaction between MAI andM+, the binding of unopsonized bacteria. We show that MAIinteract with a protease-sensitive receptor restricted to phago-cytes. The receptor is mobile in the plane of the membrane,does not require divalent cations, and mediates very rapidattachment.

MATERIALS AND METHODS

Reagents. Reagents were obtained from the followingsources: phorbol myristate acetate (PMA), fluorescein iso-thiocyanate, Tris, and poly-L-lysine, Sigma Chemical Co.(St. Louis, Mo.); trypsin, Millipore Corp. (Freehold, N.J.);Dubos base and Middlebrook 7H10, Difco Laboratories(Detroit, Mich.); Tween 20, Bio-Rad Laboratories (Rich-mond, Calif.); and neuraminidase (Vibrio cholerae), Boehr-

* Corresponding author.

inger Diagnostics (La Jolla, Calif.). Terasaki plates were

purchased from Miles Scientific, Div. Miles LaboratoriesInc. (Naperville, Ill.). Sheep erythrocytes were from bloodcollected from sheep kept at the Laboratory Animal Re-search Center of the Rockefeller University. HAP bufferconsists of Dulbecco phosphate-buffered saline (PBS; 137mM NaCl, 2.7 mM KCl, 0.9 mM CaCl2, 0.5 mM MgCl2, 8mM phosphate, pH 7.4) containing 3 mM glucose, 0.5 mg ofhuman serum albumin (Armor Pharmaceuticals, Kankakee,Ill.) per ml, and 0.3 U of aprotinin (Sigma) per ml. PD refersto PBS without calcium or magnesium.Growth and labeling of mycobacteria. MAI 13528-1079 var.

4 was provided by Anna Tsang, National Jewish Hospital,Denver, Colo. Isolates were cultured in Dubos broth con-

taining 1% Tween at 37°C in 5% carbon dioxide for 7 to 14days. Once each day, cultures were vigorously agitated byVortex mixing. MAI were harvested by centrifugation at2,000 rpm for 10 min. Cultures that were to be used withoutlabel were suspended in PBS. Cultures that were to belabeled with fluorescein were suspended in 1 mg of fluores-cein isothiocyanate per ml-50 mM sodium carbonate-100mM sodium chloride, pH 9.2, incubated at room temperaturefor 30 min, and then washed three times in PBS. The finalsuspension was sonicated for 4 s, and the optical density at600 nm was determined. Colony counts were performed byplating 15 ,ul of serial 1:10 dilutions on Middlebrook 7H10and culturing as described above. A suspension of mycobac-teria with an optical density at 600 nm of 0.2 yielded a colonycount of 5 x 109. Viability was not altered by the fluoresce-ination procedure.

Phagocytes. Human buffy coats were obtained from theNew York or San Diego Blood Bank. Monocytes were

isolated by sequential Ficoll-Hypaque gradient and Percollgradient (23). For the experiments shown in Fig. 1, 2, 3, and5, contaminating T lymphocytes were further depleted byrosetting, using neuraminidase-treated sheep erythrocytes(SRBC). SRBC previously incubated at 37°C in neuramini-dase were incubated with monocytes, and the resulting

2951

Vol. 58, No. 9

2952 CATANZARO AND WRIGHT

rosetted T lymphocytes were removed by centrifugation ona cushion of Ficoll-Hypaque. Monocytes were suspended in12.5% human serum in RPMI 1640 and cultured for 4 to 10days in Teflon beakers. These cells are hereafter referred toas M+. Polymorphonuclear (PMN) cells were prepared bythe method of English and Anderson (4) and used the sameday.

Assay for attachment of MAI to phagocytes. M( harvestedfrom Teflon beakers after 4 to 7 days were washed andsuspended at 0.5 x 106 to 1.0 x 106/ml in HAP. A 5-,ulsample of M4X suspension was added to each well of aTerasaki plate, and cells were allowed to spread for 45 min at37°C. In other experiments, monocytes or PMN were col-lected from fresh blood as described above and plated in asimilar manner. Monocytes were used at 106/ml, and PMNwere used at 2 x 106/ml. Monolayers were washed withPBS, and MAI were added at a ratio of approximately 1,000per M+. These conditions allow a synchronous, heavychallenge to the M+, and binding under these conditionsreflects the ability of M4) to bind bacteria, not the availabilityof bacteria. Plates were then incubated for 5 to 90 min at 37or 4°C. Wells were washed vigorously three times withcopious amounts ofPBS and fixed with 2.5% glutaraldehyde.The attachment of MAI was scored by fluorescence micros-copy. Phagocytes in a field were identified and counted, thelight was reduced, and UV light turned on. The number ofphagocytes with fluorescent MAI was determined. Thepercent phagocytes with at least one MAI attached wasenumerated. In addition, the number of MAI that were cellassociated was counted. Attachment index (AT) is reportedas the number of MAI that were bound per 100 phagocytes.Since each M4 bound more than 30 bacteria, enumerationwas difficult. In all experiments shown, enumeration wasfacilitated by adding a mixture of labeled and unlabeledbacteria in a ratio of 1:7. The AT determined for fluorescentlylabeled MAI was then multiplied by the ratio of labeled tounlabeled bacteria to yield the true AT. In some experiments,

monolayers of M4) were formed on a carpet of immobilizedMAI. Plastic wells were first treated with 1 mg of poly-L-lysine per ml for 30 min and washed. Various concentrationsof MAI were added and allowed to adhere for 30 min.Microscopic examination of the washed wells revealed aneven coating of bacteria at various densities. MX wereadded, and monolayers were established on these surfacesas described above.An alternative assay for attachment of MAI to MP used

cells in suspension. A 100-,ul sample of MX (106/ml) wasmixed with 100 1.l of labeled MAI an optical density of 0.2 ina plastic tube (12 by 75 mm) and incubated for 20 min at 4°Con a rotating platform. Cells were washed twice (1,000 rpm,5 min) to remove the bulk of unattached MAI, and M4) wereanalyzed in a FACScan flow cytometer (Becton Dickinsonand Co., Oxnard, Calif.).

RESULTS

Unopsonized MAI bind to M+. MAI were labeled, addedto monolayers of M4, and incubated at 37°C in the absenceof serum. Microscopic observation showed that M4 boundmore than 30 MAI during a 45-min incubation. This obser-vation was confirmed by using a suspension assay of binding(data not shown). Essentially all MO bound MAI, suggestingthat the capacity to recognize these bacteria was shared byall members of the M4 population. Some images showedMAI bound to the cell periphery, but others showed the MAIgathered in a perinuclear region, suggesting that phagocyto-sis had occurred. Since the intracellular or extracellularlocation of MAI could not be determined with certainty, wecounted all cell-associated organisms and report the data asAT.To reduce the number of labeled MAI per M4P to a level

that could be easily enumerated, a lower number of labeledMAI were mixed with a compensating number of unlabeledMAI before presentation to the MX (Fig. 1). This approach

.4:

kV r.4 _W-_FIG. 1. Photomicrograph of an M4 monolayer incubated with MAI. A monolayer of M40 was established, and a mixture of labeled and

nonlabeled MAI was added for 30 min at 37°C. Each panel shows the same field under phase-contrast illumination (A) or epillumination forfluorescence (B). The white arrow points to three labeled MAI that appear to have been ingested. Lymphocytes in the field did not bind MAI(black arrow).

INFECT. IMMUN.

BINDING OF M. AVIUM TO HUMAN LEUKOCYTES 2953

3000 [

xa)

co60)E-coC0-)

20001

1i000

10 20 30

MinutesFIG. 2. Time course of binding MAI to M+. Monolayers of M4 were established, and labeled MAI were added. After incubation of the

plates at 4°C (O) or 37°C (M) for the times indicated, Al was measured. Each point is the mean ± standard deviation of four experiments.Cells held at 4°C showed no indication of phagocytosis, but those held at 37°C appeared to ingest many MAI (see text).

lowered the number of labeled bacteria within MN4 withoutaltering the frequency of contact between phagocyte andbacteria. Labeled and unlabeled bacteria appeared to inter-act with MO by the same mechanism, since the binding of afixed concentration of labeled bacteria was competitivelyinhibited by the addition of unlabeled bacteria (data notshown).

Binding of MAI to M4 occurred very rapidly at 37°C,reaching completion by 5 min (Fig. 2). Reduction of thetemperature to 4°C slowed both the rate and the extent ofbinding about threefold. The decrease in extent of bindingcaused by lowered temperature may have been due to theinhibition of phagocytosis, since microscopic examination ofthe cultures held at 4°C showed no bacteria in a perinuclearlocation (data not shown). Binding of MAI at low tempera-ture, however, was avid, and the attachment mechanismmust be considered relatively temperature independent.

Unopsonized MAI are recognized by a proteinaceous recep-tor specific to phagocytes. Treatment of M4) with trypsinreduced the binding of MAI in a dose-dependent fashion(Fig. 3). This result suggests that a protein on the surface ofMX serves as a receptor for MAI or that the receptor isassociated with a protein. To determine the cellular distri-bution of this receptor, we measured the binding of MAI tomonolayers of monocytes, macrophages, lymphocytes, andPMN. Both monocytes and Mf bound MAI avidly, butneither lymphocytes nor PMN bound appreciable numbersof MAI (Fig. 4). The expression of receptors on the surfaceofPMN can be dramatically altered by exposure of the cellsto PMA (21). We observed that PMA treatment of PMNenabled the cells to bind MAI. These data suggest that thereceptor for MAI is restricted to phagocytes and is inducibleon PMN.

Receptors for MAI are mobile in the plane of the mem-brane. When M( are spread on surfaces coated with immu-noglobulin G or complement, the respective receptors forthese two ligands diffuse to the adherent surface of the M4and are trapped by interaction with ligand. This leaves theapical surface of the phagocyte specifically depleted of Fc orcomplement receptors (19, 22). To determine whether MAI

are recognized by a mobile receptor, M4X were allowed tospread on surfaces coated with MAI, and the attachment oflabeled MAI to the apical surface was measured. Surface-bound MAI caused strong, dose-dependent down modula-tion of the binding of bacteria to the apical portion of the M4(Fig. 5). These observations confirm that MAI are recog-nized by a specific receptor and indicate that the receptor ismobile in the plane of the membrane.

Binding of MAI does not require divalent cations. M4) bindunopsonized Escherichia coli (20), Histoplasma capsulatum

1500

xa)

4-Ic

E.-C

4-.-

1000

500

0

,ug TrypsinFIG. 3. Demonstration that treatment of M4 with trypsin inhibits

binding of MAI. An established monolayer of MX was treated with0, 1, 10, or 100 ,ug of trypsin per ml of PBS for 15 min at 37°C.Trypsin was removed by washing with HAP, labeled MAI wereadded and incubated at 4°C for 20 min, and Al was determined.Trypsin inactivated by incubation with soybean trypsin inhibitor didnot affect MAI binding (data not shown).

VOL. 58, 1990

-;F

I IujE0

2954 CATANZARO AND WRIGHT

2000 FxU1)

')cJcuE

Cr_4-'-

1i000 [

MO Monocyte Lympho- PMN PMNcyte +

PMAFIG. 4. Binding of MAI to leukocytes. Monolayers of monocytes, M+, lymphocytes, or PMN were established as described in Materials

'and Methods, MAI were added and incubated at 4°C for 20 min, and Al was determined. Each bar is the mean + standard deviation of fourexperiments. Monocytes and PMN were from freshly drawn blood; MX were from day 5 cultures. The lymphocytes observed were

contaminants in the M4 culture. Where indicated, PMN were stimulated with PMA (30 ng/ml) for 5 min at 37°C before assay.

(2), Staphylococcus aureus (6), and Leishmania species (11,17) via the CD18 complex of leukocyte antigens. All knownbinding functions of CD18 molecules are inhibited in theabsence of divalent cations. To determine whether thisfamily of receptors plays a role in recognition of MAI, we

measured binding of bacteria in the presence and absence ofdivalent cations.Monolayers of M4) were established and washed with 1

mM EDTA in PD. Monolayers were replenished with PD orwith PD and 0.5 mM CaCl2 or 0.5 mM MgCl2 or both.

80

70

60 -

2 50 -

0D 40 -

3CZ

3033

20

Labeled MAI were also washed in 1 mM EDTA in PD andthen washed in PD before being added to M+. Al wasmeasured after incubation at 4°C for 20 min. AI in eachbuffer was as follows: PD with Ca2' and Mg2+, 3,850; PDwith Ca2+, 4,142; PD with Mg2+, 4,378; PD alone, 3,259. AIswere comparable in each population and comparable tothose obtained with cells maintained in PBS throughout theexperiment (data not shown). The ability of M4i to bindEC3bi was used as a control to demonstrate that the condi-tions of the experiment functionally removed Ca2+ and

0 10 20 30

Concentration of MAI used to Coat Substrate (1 09/mI)FIG. 5. Mobility of binding sites for MAI in the plane of the membrane. Increasing concentrations of MAI were attached to the culture

surface, using poly-L-lysine as described in Materials and Methods, and a monolayer of M4 was established on the carpet of MAI. Bindingof labeled MAI to the M4, was then measured after incubation at 4°C for 20 min.

a~~~X

INFECT. IMMUN.

BINDING OF M. AVIUM TO HUMAN LEUKOCYTES 2955

Mg2". In these controls, MX binding of EC3bi was inhibited(data not shown).

Binding of MAI was unaffected by the removal of Ca2+,Mg2+, or Ca2+ and Mg2+ from the incubation medium. TheCD18 complex is thus unlikely to play a critical role in therecognition of unopsonized MAI by M+.

DISCUSSION

MAI are intracellular parasites of macrophages, and herewe show that unopsonized MAI show strong affinity formonocytes and macrophages in vitro. This property mayallow rapid colonization of appropriate host cells duringinfection. The opsonins immunoglobulin G and complementmay also serve to target MAI to M( (14), but our studiessuggest that opsonins are not necessary for the observedtropism of MAI.

Several observations suggest that unopsonized MAI arerecognized by a specific receptor protein on M+. Thebacteria are not bound by lymphocytes or inactivated PMN,ruling out a nonspecific interaction with cell membranes.Binding is also abolished by trypsinization of the M+,suggesting that a protein is involved in recognition. Finally,the capacity of M( to bind MAI is down modulated uponspreading of cells onto immobilized carpets of MAI. Byanalogy with results for Fc receptor, CR1, and CR3, thisfinding suggests that the binding sites for MAI can diffuse inthe membrane and be depleted from the apical surface byredistribution to the basal surface of the phagocyte. Thisobservation indicates that the receptors for MAI are mobileand further confirms that bulk components of the membrane,such as phospholipids, do not serve as binding sites for MAI.Members of the CD18 family of antigens on M4 serve as

receptors for several intracellular parasites. However, CD18molecules are unlikely to play a role in the binding of MAI.CD18 function is inhibited in the absence of divalent cationsand 4°C (18, 20), but these conditions do not block binding ofMAI. Further, depletion of CD18 molecules from the apicalsurface of M4 by spreading on anti-CD18 monoclonal anti-bodies does not inhibit binding of MAI (unpublished obser-vations). The mannosyl-fucosyl receptor is also unlikely toserve as the receptor for MAI, since this receptor is notexpressed on monocytes (13), a cell type that avidly bindsMAI. The nature of the receptor is now under study.

Phagocytosis is controlled and modulated by receptors onthe phagocyte. We have demonstrated that MAI are boundavidly to MX by a receptor. The avidity of binding of MAI byM4 may account for the fact that histologic examination oftissue from infected individuals showed the majority oforganisms to be intracellular (24). Although our studies showstrong binding of MAI to M+, they do not rule out theexpression of binding sites on other cell types. Indeed,Mapother and Songer (8) showed that human intestinalepithelial cell monolayers bind MAI. This finding may ex-plain the common intestinal infestation of AIDS patientswith MAI (12, 16).

ACKNOWLEDGMENTS

This study was supported by Public Health Service grants Al22003 and Al 24775 from the National Institutes of Health. S.D.W.is an Established Investigator of the American Heart Association.

LITERATURE CITED1. Barksdale, L., and K.-S. Kim. 1977. Mycobacterium. Bacteriol.

Rev. 41:217-372.2. Bullock, W. E., and S. D. Wright. 1987. Role of the adherence-

promoting receptors, CR3, LFA-1, and p150,95, in binding ofHistoplasma capsulatum by human macrophages. J. Exp. Med.165:195-210.

3. Crowle, A. J., A. Y. Tsang, A. E. Vatter, and M. H. May. 1986.Comparison of 15 laboratory and patient-derived strains ofMycobacterium avium for ability to infect and multiply incultured human macrophages. J. Clin. Microbiol. 24:812-821.

4. English, D., and B. R. Anderson. 1974. Single-step separation ofred blood cells, granulocytes, and mononuclear phagocytes ondiscontinuous density gradients of Ficoll-Hypaque. J. Immunol.Methods 5:249-252.

5. Gangadharam, P. R. J., and P. F. Pratt. 1983. In vitro responseof murine alveolar and peritoneal macrophages to Mycobacte-rium intracellulare. Am. Rev. Respir. Dis. 128:1044-1047.

6. Gordon, D. L., and J. L. Rice. 1988. Opsonin-dependent andindependent surface phagocytosis of S. aureus proceeds inde-pendently of complement and complement receptors. Immunol-ogy 64:709-714.

7. Hawkins, C. C., J. W. M. Gold, E. Whimby, T. E. Kiehn, P.Brannon, R. Cammarata, A. E. Brown, and D. Armstrong. 1986.Mycobacterium avium complex infections in patients with ac-quired immunodeficiency syndrome. Ann. Intern. Med. 105:184-188.

8. Mapother, M. E., and J. G. Songer. 1984. In vitro interaction ofMycobacterium avium with intestinal epithelial cells. Infect.Immun. 45:67-73.

9. Moulder, J. W. 1985. Comparative biology of intracellularparasitism. Microbiol. Rev. 49:299-337.

10. Rosenswieg, D. Y. 1979. Pulmonary mycobacterial infectionsdue to MAI. Chest 75:115-119.

11. Russell, D. G., and S. D. Wright. 1988. Complement receptortype 3 (CR3) binds to an arg-gly-asp containing region of themajor surface glycoprotein, gp63, of leishmania promastigotes.J. Exp. Med. 168:279-292.

12. Schmeeboum, C. W., D. M. Novick, A. B. Chabon, N. Strutyn-sky, S. R. Yancovitz, and S. Freund. 1987. Terminal ileitisassociated with Mycobacterium avium-intracellulare infectionin a homosexual man with acquired immune deficiency syn-drome. Gastroenterology 92:1127-1132.

13. Speert, D. P., and S. C. Silverstein. 1985. Phagocytosis ofunopsonized zymosan by human monocyte-derived macro-phages: maturation and inhibition by mannan. J. LeukocyteBiol. 38:655-658.

14. Swartz, R. P., D. Naai, C. W. Vogel, and H. Yeager, Jr. 1988.Differences in uptake of mycobacteria by human monocytes: arole for complement. Infect. Immun. 56:2223-2227.

15. Wolinski, E. 1979. Nontuberculous mycobacteria and associ-ated diseases. Am. Rev. Respir. Dis. 119:107-113.

16. Wong, B., F. F. Edwards, T. E. Kiehn, E. Whimbey, H.Donnelly, E. M. Bernard, J. W. M. Gold, and D. Armstrong.1985. Continuous high grade Mycobacterium avium-intracellu-lare bacteremia in patients with the acquired immune deficiencysyndrome. Am. J. Med. 78:35-40.

17. Wozencraft, A. O., G. Sayers, and J. M. Blackwell. 1986.Macrophage type 3 complement receptors mediate serum-inde-pendent binding of Leishmania donovani. J. Exp. Med. 164:1332-1337.

18. Wright, S. D., and P. A. Detmers. 1988. Adhesion-promotingreceptors of phagocytes. J. Cell Sci. 9:99-120.

19. Wright, S. D., P. A. Detmers, M. T. Jong, and B. C. Meyer.1986. Interferon-gamma depresses binding of ligand by C3b andC3bi receptors on cultured human monocytes, and effect re-versed by fibronectin. J. Exp. Med. 163:1245-1259.

20. Wright, S. D., and M. T. C. Jong. 1986. Adhesion-promotingreceptors on human macrophages recognize E. coli by bindinglipopolysaccharide. J. Exp. Med. 164:1876-1888.

21. Wright, S. D., and B. C. Meyer. 1986. Phorbol esters causesequential activation and deactivation of complement recep-tors on polymorphonuclear leukocytes. J. Immunol. 136:1759-1764.

22. Wright, S. D., P. E. Rao, W. E. Van Voorhis, L. S. Craigmyle,K. Iida, M. A. Talle, G. Goldstein, and S. C. Silverstein. 1983.Identification of the C3bi receptor on human monocytes and

VOL. 58, 1990

2956 CATANZARO AND WRIGHT

macrophages by using monoclonal antibodies. Proc. Natil.Acad. Sci. USA 80:5699-5703.

23. Wright, S. D., and S. C. Silverstein. 1982. Tumor promotingphorbol esters stimulate C3b and C3b' receptor-mediatedphagocytosis in cultured human monocytes. J. Exp. Med.156:1149-1164.

INFECT. IMMUN.

24. Young, L. S. 1988. Mycobacterium avium complex infection. J.Infect. Dis. 157:863-868.

25. Young, L. S., C. B. Inderlied, 0. G. Berlin, and M. S. Gotlieb.1986. Mycobacterial infections in AIDS patients, with an em-phasis on the Mycobacterium avium complex. Rev. Infect. Dis.8:1024-1033.